X
Tech

Handwriting recognition that really works

Tired of having to adjust to the quirks of handwriting recognition systems? This one from Sweden lets you write naturally--without any special symbols or arcane strokes to remember.
Written by Rich Castagna, Contributor
The message is indeed the medium, but sometimes this concept gets lost in the exuberance of the breakneck pace of technology development.

Whatever the form of your favorite computer--whether it sits on your knee, in the palm of your hand, or even clips onto your clothing somewhere--and regardless of all the bells and whistles it might offer, its bottom line function is managing information.

You need to store, send and receive, and create data. Computer manufacturers successfully tackled the storing and transmission functions years ago, as evidenced by today's mobile computers that let you carry reams of information in a tiny device that can also keep you in touch via e-mail and voice communications.

In retrospect, you could say that was the easy part. The toughest nut to crack is the creation side of the mobile computing equation. While a wealth of work has been done on refining the three principle input modes--voice, handwriting, and keyboarding--no single technology has risen to the top. The slimmed-down, folding keyboards, like Pocketop's versatile product, are probably the best mobile input alternatives right now if you need to enter anything more than a few words. But these, too, have some drawbacks, most notably the need to stake out at least a modest piece of desktop real estate to use them effectively.

Handwriting is natural, quick, and can be done in just about any position. But to date, even the best of the handwriting recognition systems require painstaking adherence to the software's idiosyncrasies, which usually means abandoning anything close to your natural handwriting style. Decuma AB, a Swedish company, thinks it has a solid solution to the handwritten input dilemma with its OnSpot recognition system.

OnSpot is the most accurate handwriting system I've ever seen. It doesn't yet handle cursive writing, so you'll stick to printed characters, but its accuracy is excellent. What's likely to strike you first about OnSpot is its patience--it seemingly knows just how long to wait for you to complete a character. So, if you tend to write a "t" with two discreet strokes, the software will correctly recognize that the two strokes are elements of a single character--unlike most other recognition systems that turn the first vertical stroke into an "i." You can even add accents, umlauts, and other special symbols to the characters you write.

But OnSpot is fast, too, so you can print at a comfortable, natural pace. It's also incredibly easy to correct mistakes--either yours or misinterpreted characters: You just write the correct character over the wrong one--no special strokes to delete or backspace are necessary. Another useful feature is OnSpot's ability to store text that you can recall; for example, you store a greeting that you frequently use for your e-mail correspondence and associate it with a simple graphic like a smiley face. To recall it, you just draw the graphic again.

Yet another OnSpot feature relieves an annoyance familiar to anyone who handwrites data on a PDA. The input areas on most PDAs are fairly limited, and with most recognition systems you have to be conscious of when you're about to run out of space so that you can transfer the text and get a clean slate to continue writing. OnSpot offers a simple, yet effective solution: As you approach the edge of the writing area, you just put a hyphen in the word; the program will move your text and let you complete the hyphenated word. Not a big deal maybe, but it certainly helps speed things up.

There several versions of OnSpot--one for English and European languages, a Chinese version, and an edition that recognizes Japanese characters.

OnSpot isn't available as a separate product at this time, as it's being marketed to mobile hardware and software vendors. It's so good that it's likely to start showing up in plenty of PDAs and can probably even improve on the recognition systems in the new crop of tablet computers.

Is your PDA's current handwriting recognition system good enough? TalkBack below or e-mail us.

Editorial standards